The Effects of Discourse Markers On The Reading Comprehension and Speed of Chinese Learners of English

International Journal of English Language and Linguistics Studies
Vol.2, No.2, pp.27-49, June 2014

Published by European Centre for Research Training and Development UK (www.ea-journals.org)
THE EFFECTS OF DISCOURSE MARKERS ON THE READING

COMPREHENSION AND SPEED OF CHINESE LEARNERS OF ENGLISH
Zhang Ang
Intensive Language Training Center
Sichuan University
No. 24, South District 1, the 1st Ring Road
Chengdu, Sichuan
Peoples Republic of China
ABSTRACT: Discourse markers are frequently used in academic papers, which abound in
logical analyses and arguments. Some linguistic theories have shown that discourse markers
are indispensable for both cohesion and coherence of these papers. In this study, a controlled
test was carried out to analyze the way in which discourse markers affect the reading
comprehension and reading speed of Chinese learners of English. In the study two groups of
students (undergraduates and postgraduates) who minored in English as a foreign language
were tested with four versions of a medical paper. Four versions were designed: a version
without discourse markers, a version with micro markers, a version with macro markers, and
a version with both micro and macro markers. A cloze recall test was used to measure the
students comprehension and reading speed. The results show that macro markers play a
helpful role in enhancing readers reading comprehension and reading speed. It is concluded
that Chinese teachers should place more emphasis on the instruction of discourse markers,
especially macro markers.
KEYWORDS: Discourse Markers, Reading Comprehension, Reading Speed, Chinese
Learners
INTRODUCTION
With increasing research on how to improve the reading comprehension as well as the reading
speed of Chinese learners of English, the roles of discourse markers have been extensively
studied. The first monograph entitled Discourse Markers (Dr. Schiffrin, reprinted in 1996) was
published in 1987. The author made a comparative analysis of markers within conversational
discourse1 collected during sociolinguistic fieldwork. Dr. Schiffrin pointed out that discourse
markers perform important functions in conversation because they provide contextual
coordinates which aid in the production and interpretation of coherent conversations at both
local and global levels of organization. However, Dr. Schiffrin focused on the functions of
discourse markers in conversations and not in reading. Moreover, the emphasis of her work
was merely put on English as a first language, rather than English as a second or foreign
language.
See Notes
27

Vol.2, No.2, pp.27-49, June 2014
Dr. Chaudron (19462006) was the first scholar who studied the roles played by discourses
markers in enhancing the comprehensions of lectures. In his paper (Dr. Chaudron, 1986) Dr.
Chaudron surveyed the way in which different categories of discourse markers affect foreign
students understanding of university lectures given in English. He classified discourse markers
into two categories: micro-markers such as BUT, SO, etc. and macro-markers such as Lets
go back to the beginning. In order to study the functions of these markers, he designed a
survey, in which four versions of the same lecture material were designed: the versions with
discourse markers and the version without discourse markers. Then the learners of English as
a second language at different proficiency levels were asked to listen to these versions of
lecture. The results show that the Macro version helped learners best.
In 2009, Dr. Martinez of Department of English Philology, Faculty of Philology University of
Oviedo studied the effects of discourse markers on the reading comprehension of Spanish
students of English as a foreign language. (Dr. Ana Cristin Lahuerta Martinez, 2009)
According to Ana, she has found a significant correlation between the presence of discourse
makers in the text and reading comprehension. Her study did not cover whether discourse
markers affected the reading speed. This paper attempts to survey to what extent different
categories of discourse markers affect both the reading comprehension and the reading speed
of Chinese learners of English.
SOME THEORIES
Grammar Analysis vs. Discourse Analysis
It is a common phenomenon that Chinese teachers of English are used to focusing their
attention on grammar analyses while teaching English reading. This, in turn, leads to the fact
that many English learners find it very slow to enhance their reading comprehension and speed.
2(Dr. Yuan, 1982) It is evident today that discourse analysis should be emphasized if we want
to help students enhance both the reading comprehension and the reading speed.
Grammar analysis aims mainly at dissecting a well-formed sentence into pieces or bits o
explain its internal relationship, while discourse analysis treats the passage as a coherent whole
and tries to discover the relationship between the sentences instead of that between the parts in
a sentence. (Dr. Brown and Yule, 2003) By comparison, discourse analysis has studied how
rhetorical form or macro-structure governs the writers syntactic options, and how these in turn
help the reader to grasp rhetorical form and to understand compositions as wholes. (Dr.
Goodin, 1982)
Cohesion, Coherence and Discourse Markers
In discourse analysis, stress is laid on the understanding of cohesion and coherence of the
passage, rather than its grammatical structure. Dr. Omaggio (2000) makes a good description
of the importance of cohesion and coherenceCohesion is concerned with the way
2
See Notes
28

Vol.2, No.2, pp.27-49, June 2014
propositions are linked structurally in a text and how the literal meaning of a text is interpreted.
Cohesive elements include the use of pronouns, grammatical connectors, and lexical cohesion
and the direct repetition of the same term to refer to the same object. Dr. Omaggio described
coherence this way: Coherence refers to the ways contextual meaning are related and linked
in the parts of the discourse; that is, it refers to the relationships among the different ideas in a
text. In a text, every sentence forms part of a developing, cumulative instruction which tells
us how to construct a coherent representation. To state in a simpler way, cohesion refers to
the relationship of linguistic elements within a text, while coherence reflects the consistency of
meaning. So cohesion and coherence are not the same.Dr. Carell (1982) has studied the
relationship between cohesion and coherence and pointed out: A text will have a low level of
cohesion but still be highly coherent if it draws on a shared domain of knowledge and allows
for a high level of inference. The reverse is also true: a high level of cohesion does not ensure
coherence when the text reflects a misunderstanding of real-world events.
Lets look at the following two examples.
Example 1: John is a professor in this field. He is able to help you with this problem.
Here, without the use of a discourse marker such as And, So, and Thus, a reader can still
infer the Cause and Effect relationship between the two sentences. This is because the text
itself draws on a shared domain of knowledge: a professor is an expert in a field. Since John
is a professor in this field, he should be able to help one solve the problem. Therefore, this text
can be considered cohesive even though no discourse marker is used to connect the two
sentences.
Example 2: John is a professor in this field. And he is unable to help you with this problem.
Here, the discourse marker And is used to link the two sentences, but the text is not coherent.
Instead, it makes the reader confused. Dr. Halliday and Dr. Hasan (1976) have emphasized the
importance of cohesion of a text and the use of discourse markers (conjunctions in particular)
to achieve cohesion. They describe them this way: Thus, the concept of cohesion accounts
for the essential semantic relations whereby any passage of speech or writing is able to function
as a text. We can systematize this concept by classifying it into a small number of distinct
categories reference, substitution, ellipsis, conjunction, and lexical cohesion.
In short, discourse markers serve the purpose of cohesion and reveal explicitly the underlying
relationship between sentences in a text.
Types of Discourse Markers
According to Dr. Halliday and Dr. Hasan (1976), discourse markers can be classified into three
categories:
Coordinating conjunctions: and, but, for,
Conjunctive adverbs: furthermore, however, therefore;
Prepositional phrases: in addition; in spite of; as a result;
In terms of their functions, discourse markers are classified by Dr. Halliday and Dr.Hasan into
four types:
29

Vol.2, No.2, pp.27-49, June 2014
Additive: and, furthermore; in addition;

Adversative: but, however; in spite of
Causal: for; so; therefore; due to;
Temporal: first; then; finally;
Dr. Quirk et al (1972, p. 664) further include some other expressions into the categories of
markers (see Table 1). And Dr. Cook (1975) names them as macro-markers because they
signal the macro-structure of a text.
Table 1: Quirks Classification of Certain Expressions
Category
expressions
Enumeration
The first point I want to make is this
Transition
I want to begin by saying
Summation
Apposition
Let us now turn to

The next thing is
Result
Inference
Reformation
Replacement
My conclusion is
I will sum up by saying
Another example is
This means that
Concession
The consequence was
This is because
That implies..
A better way of putting it is
The alternative is
Another possibility would be
It is true
The truth is that
Roles of Discourse Markers in Discourse Analysis
Discourse markers make a text cohesive by promoting clarity. Just as Dr. Hans Guth points
out: Apt transitional phrases help the reader move smoothly from one point to the next. (Dr.
Guth, 1980, p.49) Also, as Sloan suggests, In order to avoid the unclarity of the discourse,
discourse markers must be used, especially in scientific papers which are characterized by so
many logical analyses and arguments. (Dr. Sloan, 1986, p.168) Hence in scientific papers,
readers can often see various discourse markers in almost each and every paragraph.
However, macro-markers differ from micro- markers in their features and effects. Macromarkers help the reader better understand the macrostructure of a text by revealing the major
information contained in the text and the arrangement of that information, whereas micromarkers are to assist the reader in discerning the links between sentences within a text. As
30

Vol.2, No.2, pp.27-49, June 2014
micro-markers help to reveal the internal relationship within transitional units or between
sentences, the reader will find it easy to grasp the discourse efficiently. However, even though
relational signals sometimes lie within transitional units already, micro-markers are still
needed. Under certain circumstances the absence or omission of micro-markers may create
ambiguity that can lead the reader to the misunderstanding or misreading. Two examples are
cited below to explain the importance of micro-markers.
Example 1:
"It is hard to answer that question with reference to a short time period because no one has
really studies it. BUT for long-term studies, ductal ligation is a very important model for
inducing pancreatic atrophy."
If we take out " BUT ", the relationship between the two sentences will be non-sequential, and
as a result, ambiguity occurs.
Example 2:
"He was an unqualified doctor. HOWEVER, he was assigned to study medicine at a medical
university for three years."
If "However" is removed, readers may consider the relationship between the two sentences as
"cause and effect", and then they may think that the doctor was thus sent to get more training
at a medical university because he was unqualified at the moment. In fact, this sentence means
that though the doctor was unqualified, others had to trust him with the thought that he studied
medicine in any case.
The second rule of micro-markers is to act as fillers to allow readers to have a pause and to
catch their breaths before plunging into the next thought. As Sloan says, "The written language,
an outgrowth of speech, should also allow far the momentary suspension of thought. ( Dr.
Sloan, 1986, p.175)
On the other hand, Dr. Schlffrin suggests, "Cohesive devices do not themselves create meaning;
they are clues used by speakers and hearers to find the meanings which underlie surface
utterances. (Dr. Schiffrin 1988, p.9) And discourse markers are the most commonly used
cohesive devices.
Then, how do readers manage to use these clues to find the meanings underlying surface
writing? The next section discusses relationships between reading comprehension processes
and the use of discourse markers.
Schema Theory
Based on the study of cognitive psychology, the schema theory is mainly concerned with how
various types of background knowledge affect the understanding of a text. Schema is referred
to as the building block of cognition which is stored hierarchically in one's long-term
memory.(Dr. Rumelhart, 1981) Schema in people's minds can be of various types. As we all
know, during cognition of some new information, one always associates a piece of new
information with old things or experiences which have already been stored in ones mind. In
other words, the cognition or comprehension of any new information is very often influenced
or controlled by the schemata existing in ones mind. The incoming new information must be
related to the old schemata. Schema can not only give one some guidance to comprehending
31

Vol.2, No.2, pp.27-49, June 2014
all sorts of new things and experiences, but also help one describe these things and experiences
in words.
According to Dr. Carrell (1982), scheme can be further classified into two types: (1) formal
schema, which reflects ones background knowledge about the rhetorical or discourse
structures of different types of texts; and (2) content schema, which refers to one's knowledge
about the external world, objects, events, and situation as well as one's expectations. For
example, while the reader is reading a text, he usually employs his background knowledge
about the real-world to expect what the text is going to talk about. So, content schema will
guide his comprehension of events, objects, and situation. Moreover, content schema is more
important if the relationship between sentences is not illustrated in an explicit way. In this case,
it is left to the reader to understand the topic and to decode the underlying relationship between
sentences, which has to rely upon the functions of content schema.
Let's look at the following examples:
"The young Chinese mother was so glad to finally give birth to a son, a second child. Three
days later, she was dismissed. "
Even with a close reading of the sentences, a native English speaker will be at a loss if he is
unaware of China's one-child policy. But a Chinese reader can easily understand that the mother
was dismissed as a punishment for her illegal second delivery. She had a second child at the
cost of losing her job. Here we can see how content schema has its influence on ones reading
comprehension. At the same time, the reader while reading also uses "formal schema", i.e. his
knowledge of the rhetorical structure of the discourse, which directs him to correctly interpret
the linguistic representation of the events and activities described in the text. While reading a
paper, for example, those who are well-acquainted with the discourse structure of the
introduction part of the paper may have no difficulty reading the whole paper.
Dr. Kintsch and van Dijk made a good summary of the functions of the schema. They claim,
"It is the schema which determines which of the many propositions in a text are relevant or
irrelevant to the reader, and thus directly affects how and whether they are processed or
recalled. If the reader has a schema that is not well defined, the outcome of the processing of
the text will be haphazard with obvious problems for comprehension, whereas if he is familiar
with conventional nature of the text, well-defined schemata will be produced, which will aid
comprehension and recall. (Dr. Kintsch and van Dijk, 1975) Based on this theory, if a reader,
especially a reader whose first language is not English, is unfamiliar with conventional nature
of a text, well-defined schemata will not be produced, which can not help comprehension and
recall. With this in mind, the use of discourse markers should play a good role in aiding a
reader comprehend and recall well.
The objective of this paper is to analyze the way in which readers react to discourse markers in
a text.
32

Vol.2, No.2, pp.27-49, June 2014
PURPOSE and HYPOTHESES

The purpose of the present survey is to answer two questions concerning reading academic
papers.
(1) What are the effects of macro-markers on readers reading comprehension as well as
reading speed?
(2) What are the effects of micro-markers on readers reading comprehension as well as reading
speed?
The survey was based on a paper on medical ethics chosen from "Journal of Medical Ethics".
Four different versions were designed for the same paper. They were used to test
undergraduates and graduates who minored in English. Then, measurements of their reading
comprehension and speed were taken.
The hypotheses of the survey are that the more discourse markers are used, the better readers
comprehend the paper and the faster they read it. In other words, with more discourse markers
used in the paper, readers may catch hold of more information conveyed in the paper and read
more smoothly.
More specifically, the hypotheses are:
1. If the reading speeds of the investigated subjects are the same, then:
(1) they would comprehend the paper with macro-markers better than the paper without any
discourse markers.
(2) they would comprehend the paper with micro-markers better than the paper with only
macro- markers.
(3) they would comprehend the paper with both micro-markers and macro-markers best.
2. If the reading comprehension of the subjects is the same, then:
(1) they would read the paper with macro-markers faster than the paper without any discourse
markers.
(2) they would read the paper with micro-markers faster than the paper with only macromarkers.
(3) they would read the paper with both micro-markers and macro-markers fastest.
MATERIALS
A paper read by Dr. Jean Davies (1988) and published in an issue of Journal of Medical
Ethics ( See Appendix C ) was chosen as the source material. This paper dealt with the
problem of euthanasia in Great Britain. As the topic of this sort had been widely discussed in
newspapers and journals, the contents of the paper were supposed to be familiar to medical
students in China. Moreover, the writing style of the paper was not difficult for university
students to understand. Hence, both the contents and the writing style of the paper would not
cause the reader any hindrance in reading and understanding, which guaranteed that
background knowledge and difficulty of the subject would not affect this experiment. The first
version of the paper, i.e. the Baseline Version, contained no discourse markers. However,
caution was given to avoid any influence of conveying the meaning as the result of lack of
discourse markers.
33

Vol.2, No.2, pp.27-49, June 2014
In the second version of the paper, i.e. the Micro Version, only micro-markers were employed.
But on other changes were made concerning the contents of the paper. These micro-markers
used were mainly as follows (for the complete list of the discourse markers used, see Appendix
A):
ADDITIVE:
moreover, in addition;
APPOSITION: for example;
COMPARISON: similarly;
ADVERSATIVE: however, nevertheless;
CONTRACTIVE: in fact;
GENERAL CAUSAL: so, then, therefore;
Attempts were made not to allow any of these discourse markers to add any semantic meaning
to the paper. The semantic relationships marked by the micro-markers had already underlined
the text. These markers gave no new propositional meaning to the text.
The third version, called the Macro Version, contained signals about the major propositions to
show certain important transition points. The following are some examples (for the complete
list, see Appendix A).
ADDITIVE: What is more,...
APPOSITION: Another example is that...
CAUSE:
This is because...
SUMMATION: It is concluded that...
Also, these markers gave no new propositional meaning to the text.
In the fourth version, called the Micro-Macro Version, both micro and macro markers were
employed.
The following are some examples to show the four different versions (for the complete paper
version, see Appendix B):
Baseline Version:
The deliberate killing of one private citizen by another is deeply disapproved of in our society.
The action is defined as a crime.
(2) Micro Version:
Furthermore, the action is defined as a crime.
(3) Macro Version:
What is more, the action is defined as a crime.
(4) Micro-Macro Version:
And what is more, the action is defined as a crime.
Since different discourse markers were used in the four different versions, their total word
numbers were different, as shown in Table 2.
Table 2: Total Word Number of Each of the Four Versions
Paper version
Baseline Version
Micro Version
Macro Version
Micro-Macro Version
Total words
1101
1139
1185
1200
34

Vol.2, No.2, pp.27-49, June 2014
These four versions were made by this writer and checked by two American colleagues. The
principles were 1) to guarantee these revised papers, especially the baseline version, can be
read smoothly; 2) not to let the difference of word count in the revised papers affect
comprehension and recall.
METHODS
The recall cloze test was chosen as the method to measure the reading comprehension of the
investigated students because this sort of test has been shown by certain scholars to be a
relatively reliable measurement of the reader's reading comprehension (Dr. Henning et al,
1981). The method used to check the reading speed was to ask the investigated subjects to
record their exact reading time they used. Before the real test was carried out, several
preparatory tests were undertaken to make sure: (1) the contents of the paper and its questions
for comprehension would not interfere with the reading comprehension; (2) the recall cloze test
used was reliable; and (3) the way of subjects recording their reading time was reliable and
accurate. The preparatory tests were done on ten graduate students and ten undergraduate
students of Xian Medical University. The former had studied English for more 10 years; the
latter for 7 years. The results showed that both groups of students could comprehend and recall
the main ideas of the papers.
Investigated Subjects
In order to find out whether discourse markers would affect the reading comprehension and
speed of both intermediate and advanced English learners, two groups of subjects were chosen
for taking the test. Group one was made up of medical undergraduates; group two was made
up of medical postgraduates. The undergraduate group consisted of 64 third-year medical
students of Xi'an Medical University. Having studied College English for two years, the
students of this group had achieved the intermediate level of English proficiency, a level which
is close to Band 5 of the IELTS test. The students of the postgraduate group were considered
at an advanced level of proficiency, a level which is close to Band 6 of the IELTS. 68 firstyear graduate students enrolled at Sichuan University were chosen at random. And they were
at that moment taking a reading comprehension course given by Chinese and American
instructors.
Procedures
The four versions of the paper were given at random to all the subjects. They were informed
of this test in advance. The test was given at their regular class time, with the presence of their
English instructor. Before the test papers were handed out, the detailed explanation for the
purpose and procedures of test was made in Chinese so as to avoid any misunderstanding by
the subjects. Then, they were asked to pay attention to the following requirements:
35

Vol.2, No.2, pp.27-49, June 2014
(1) Write down accurately the starting time before reading, and the finishing time when the
reading is over.
(2) Read the paper only once. No time is given for a second reading. Reverse the reading paper
after finishing reading it and do a recall cloze test.
(3) No note-taking is allowed.
Analysis
The score measurements and the reading-speed records were analyzed separately in an analysis
of variance (SPSS/PC )3 In addition, the comparisons of means in pairs were made according
to the above-mentioned hypotheses so as to find out the different effects of discourse markers
in different versions. For example, the Macro Version was compared with the Micro Version;
the Macro Version was with the Micro-Macro Version; the Macro Version was with the
Baseline Version; the Micro Version was with the Micro-Macro Version; the Micro Version
was with the Baseline Version, and finally, the Micro-Macro Versions was compared with the
Baseline Version.
RESULTS
The results of the two groups are reported separately as follows.
l. The undergraduate group
(1) Score measurements
Table 3 shows means of recall scores. The full score is 10 points.
Version
3
4.
Table 3: Means of Recall Scores

N4
Mean
s.d.5
SPSS/PC: Statistical Program for Social Science -- Personal Computer. See Notes
" N " refers to number of the subjects.
5. "s.d." refers to "standard deviation "

6. "SS" refers to "sum of squares".
7. "d f refers to "degree of freedom".
8. "MS" refers to "mean square ".
9. "F" refers to calculated statistics.
10. "P" refers to probability value. See notes
36

Vol.2, No.2, pp.27-49, June 2014
Baseline
Micro
Macro
Micro-Macro
Overall mean
16
17
14
17
64
3.1250
2.5294
4.1429
2.6471
3.111
1.6279
1.1246
2.3158
1.4550
Table 4: Analysis of Variance of Scores by Version

Source
SS6
d f7
MS8
F9
Total
187.75
63
Between
24.16809
3
8.05603
2.95486*
Groups
Within Groups 163.5819
60
2.726365
* P10< 0.05
Table 4 shows the results of the Analysis of Variance in this measurement as affected by the
different paper versions. It is evident from Table 4 that there is a significant effect of the version
(F3/60) = 2.95486, P<0.05). Table 5 shows the results of comparison of means in pairs. This
analysis is aimed at making comparison of differences to find out which pair of comparison
shows significant difference. From Table 5, it is noted that there appears no significant
differences between the means of the Macro and the Baseline Versions. This finding conflicts
with the hypothesis. Also, in contrast with the hypothesis, the Marco Version is superior to the
Micro Version (P<0.05), and moreover, it is even superior to the combined Micro-Macro
Version. However, the comparisons reveal no differences whatever between the Micro-Macro
Version and the Micro Version or the Baseline Version.
Table 5: Results of Comparison of Means in Pairs
Pairs
Q value11
Macro vs.
Baseline
2.5417
Macro vs.
Micro-Macro
Macro vs.
Micro
11
12
13
14
a12
v13
q
(0.05;0.01)14
P value
60
2.83; 3.76
P>0.05
3.6781
60
3.4; 4.28
0.01<P<0.07
3.829
60
3.74; 4.59
0.01<P<0.05
Q value refers to calculated statistics;

a refers to the number of groups;
v refers to degree of freedom within groups;
q refers to significant level;
37

Vol.2, No.2, pp.27-49, June 2014
Baseline vs.
Micro-Macro
1.4752
60
2.83- 3.76
P>0.05
Baseline vs.
Micro
1.4134
60
3.4; 4.28
P>0.05
Micro-Macro vs.
Micro
.2753
60
2.83;3.76
P>0.05
Reading Speed Measurement

Table 6 shows the means of reading speed for the four versions.
Table 6: Means of Reading Speed
Version
N
Mean (wpm)
Baseline
16
77.5562
Micro
17
87.8365
Macro
14
94.6929
Micro-Macro
17
82.8606
Overall
64
Overall mean
85.7366
s.d.
15.80
16.89
15.96
16.18
Table 7 shows the results of the Analysis of Variance in the reading speed measurements as
affected by the different paper versions.
Table 7: Analysis of Variance of Reading Speed by Versions
Source
SS
df
MS
F
Total
18212.13
63
Between
2403.75
3
801.25
3.04111*
Groups
Within Groups 15808.38
60
263.4729
* P15< 0.05
It is clear from Table 7 that there is a significant effect of the version on the reading speed.
Table 8 shows the results of comparison of means in pairs. The purpose is also to find out
which pair has a significant difference between means of the reading speed.
Pairs
Macro vs.
Micro
Macro vs.
Micro-Macro
Q value
q (0.05;0.01)
P value
1.7416
60
2.83; 3.76
P>0.05
2.9597
60
3.4; 4.28
P>0.05
38

Vol.2, No.2, pp.27-49, June 2014
Macro vs.
Baseline
Micro vs.
Micro-Macro
4.1369
60
3.74; 4.59
0.01<P<0.05
1.2446
60
2.83- 3.76
P>0.05
Micro vs.
Baseline
2.4817
60
3.4; 4.28
P>0.05
Micro-Macro vs.
Baseline
1.2628
60
2.83;3.76
P>0.05
Based on Table 8, the subjects could read the Macro Version faster than those who read the
Baseline Version (P<0.05). This result supports the hypothesis. However, in conflict with the
hypothesized effect, there are no differences in comparison of means between the Micro
Version and the Macro Version; nor are there any differences between the Micro-Macro
Version and the Macro Version. These comparisons also reveal no difference between the
Micro-Macro Version and the Micro Version; no difference between the Micro-Macro Version
and the Baseline Version (P>0.05).
Relationship between the score and the reading speed
Table 9 below shows that there exists a significant difference (P<0.05) in score means between
the Macro Version and the Micro Version while the reading speeds for the two versions are
similar (P>0.05). Clearly, in terms of accelerating the reading comprehension, the Macro
Version is significantly superior to the Micro Version. Also, the recall score mean for the
Macro Version is significantly higher than that for the Micro-Macro Version (P<0.05) while
the reading speeds for the two versions show no difference at all (P>0.05). In direct conflict
with the hypothesis, the combination of the Micro and Macro markers can not help the subjects
to best comprehend the paper, nor can they make them read fastest. While the score means for
the Macro and the Baseline Versions reveal no significant differences, there does exist a
significant difference between the reading speeds (P<0.05). This supports the hypothesis.
Table 9: Means of Answer Scores and Reading Speed Measurements
Version
N
Score
Speed (wpm)
Baseline
16
3.13
77.5562
Micro
17
2.59
87.8365
Macro
14
4.14
94.6929
Micro-Macro
17
2.65
82.8606
Overall
64
Overall mean
3.111
85.7366
To sum up, the findings show that the macro-markers help the subjects increase not only their
reading comprehension but also their reading speed. However, the micro-markers do not show
any superiority in this regard.
39

Vol.2, No.2, pp.27-49, June 2014
The Postgraduate Group

Table 10 shows the means of the recall measurements.
Table 10: Means of Recall Measurements
Version
N
Mean
Baseline
17
2.6471
Micro
19
3.7895
Macro
15
4.5333
Micro-Macro
17
3.6471
Overall
68
Overall mean
3.60425
s.d.
1.46
1.93
1.72
2.1
1.8
Table 11 suggests the results of the Analysis of Variance in the measurements as affected by
the different paper versions.
Table 11: Analysis of Variance of Score Means by Versions
Source
SS
df
MS
F
Total
229.8089
67
Between
29.15289
3
9.717631
3.099477*
Groups
Within Groups
200.656
64
3.135249
* P< 0.05
It can be evidently seen in Table 11 that there is a significant effect on the version in the cloze
responses ( F3/64 = 3.099477, P<0.05). Table 12 shows the results of the comparison of
means in pairs, the purpose of which is to find out which pair of comparison shows a significant
difference. From Table 12, it is noted that there are no differences in means between the Macro
Version and the Micro Version, which is contradictory to the hypothesis. Also, in conflict with
the hypothesis, the Micro-Macro Version is not superior to the Baseline Version, the Micro
Version as well as the Macro Version, in terms of facilitating the reading comprehension.
However, in agreement with the hypothesis, the Macro Version is significantly better than the
Baseline Version (P<0.05). Other results reveal no difference between the Micro Version and
the Baseline Version. Hence, it is of interest to note that, among the four versions, only the
Macro Version is significantly superior to the Baseline with its advantageous effects on the
reading comprehension, although the Micro and Micro-Macro Versions seem to produce better
results than the Baseline Version.
Pairs
Macro vs.
Micro
Macro vs.
Micro-Macro
Q value
q (0.05;0.01)
P value
1.7796
64
2.83; 3.76
P>0.05
2.1203
64
3.4; 4.28
P>0.05
40

Vol.2, No.2, pp.27-49, June 2014
Macro vs.
Baseline
Micro vs.
Micro-Macro
4.3618
64
3.74; 4.59
0.01<P<0.05
.3316
64
2.83- 3.76
P>0.05
Micro vs.
Baseline
2.5757
64
3.4; 4.28
P>0.05
Micro-Macro vs.
Baseline
2.2546
64
2.83;3.76
P>0.05
Reading Speed Measurements

Table 13 shows the means of the reading speeds for the four different versions.
Table 13: Means of Reading Speed
Version
N
Mean (wpm)
s.d.
Baseline
17
76.3947
14.17
Micro
19
83.6678
12.54
Macro
15
93.3847
16.89
Micro-Macro
17
83.8
17.56
Overall
68
Overall mean
84.3118
Table 14 shows the results of the Analysis of Variance in the reading speed measurement as
affected by the different paper versions.
Table 14: Analysis of Variance in Reading Speed by Versions
Source
SS
df
MS
Total
17281.69
67
Between
2307.063
3
769.0208
Groups
Within Groups
14974.63
64
233.9785
* P< 0.05
F
3.286816*
It is clear from Table 14 that there are significant effects of the different versions on the
reading speed.
Table 15 shows the results of comparison of means in pairs, the purpose of which is to figure
out which pair of comparison shows a significant difference in means of reading speeds.
Pairs
Macro vs.
Micro
Q value
2.6543
a
2
q (0.05;0.01)
P value
64
2.83; 3.76
P>0.05
41

Vol.2, No.2, pp.27-49, June 2014
Macro vs.
Micro-Macro
2.6909
64
3.4; 4.28
P>0.05
4.5478
64
3.74; 4.59
0.01<P<0.05
.0356
64
2.83- 3.76
P>0.05
Micro vs.
Baseline
1.9327
64
3.4; 4.28
P>0.05
Micro-Macro vs.
Baseline
1.8982
64
2.83;3.76
P>0.05
Macro vs.
Baseline
Micro vs.
Micro-Macro
Just as what the hypothesis supposes, the Macro Version is significantly superior to the
Baseline Version in augmenting the reading speed. However, in conflict with the hypotheses,
other results reveal no differences between pairs. It is worth noticing that in this comparison
only the Macro Version is significantly superior to the Baseline.
Relationship between the Scores and the Reading Speed Measurements
As is seen in Table 16, the subjects of the Micro Version could achieve nearly the same reading
speed in terms of getting the similar amount of information which the subjects of the Macro
Version could get. This finding is contradictory to the hypothesis: the subjects would
comprehend the paper with micro-markers better than the one with only macro-markers. With
the measurement supporting the hypothesis, the subjects who read the Macro Version could
read faster and acquire more information than those who read the Baseline Version. Again, in
conflict with the original hypothesis, it turned out that the subjects reading the Micro-Macro
Version could not comprehend the paper best; nor could they read it at the fastest speed.
Table 16: Means of Answer Scores and Reading Speed Measurements
Version
N
Score
Speed (wpm)
Baseline
17
2.6471
76.3947
Micro
19
3.7895
83.6679
Macro
15
4.5333
93.3849
Micro-Macro
17
3.6471
83.8
To sum up, the findings show that macro-markers not only facilitate the reading comprehension
but also increase the reading speed. By contrast, micro-markers fail to show this sort of
superiority.
DISCUSSION
The first point for discussion is why micro-markers did not help the subjects of both groups to
better understand the contents of the paper and enhance the reading speed.
42

Vol.2, No.2, pp.27-49, June 2014
For one thing, unlike macro-markers which are generally considered as the higher-order
discourse markers that signal major transitions and emphasis in the paper, micro-markers are
the lower-order markers of segmentation and intersentential connections. In this respect, they
probably do not add enough content to make the subsequent information more salient or
meaningful.And for another, with a number of micro-markers scattered throughout the paper,
this might probably make the whole paper look less well-organized, just as what is explained
by Dr. Hiller et al (1968) in their research. Dr. Sloan (1986) also points out that only 7% of all
transitional units required micro-markers to ensure clarity (here he means instant clarity). After
analyzing 25 essays, Sloan found that micro-markers are never abso1utely essential to
revealing the meaning of the essays.
Thirdly, due to the lack of the knowledge of discourse analysis and discourse markers, Chinese
learners of English have got into the habit of decoding a paper word by word, rather than
extracting the information out of the paper through comprehending the discourse devices.
Consequently, while reading the Micro Version, the subjects either neglected micro-markers
or spent more time decoding them, especially those unfamiliar micro-markers. As a result, their
reading speed slowed down.
Fourthly, because Chinese teachers of English usually focus on grammar analysis instead of on
discourse analysis in their teaching, their students can not be well-informed of the knowledge
of micro-markers and hence are inclined to overlook the transitional implication of micromarkers. As a result, based on this survey, micro-markers could not play a useful role in
enhancing the subjects efficient reading even though they came across these markers.
Furthermore, as they had to make frequent stops to think about some unfamiliar micro-markers,
this might have made them lose the thread of thinking and memorization. Thus, they did not
benefit from the use of micro markers while reading the paper with micro-markers. It seemed
to them that the Micro Version was the same as the Baseline Version.
The second point for discussion is whether it is true that, as pointed out by Dr.Sloan (1986),
the absence of some obligatory markers affects the reading speed, but it does not or seldom
affect the reading comprehension. Dr. Sloan explained in this way: "Whatever misreading
results from the absence of obligatory markers is almost always fleeting. The next sentence or
two usually correct any misconception by supplying information that restores clarity to the
momentarily puzzling."
This survey proves this theory because it explains why the subjects of the undergraduate group
who read the Baseline Version could get the same amount of information as those who read
the Macro Version at a slower speed. With the same theory, the subjects of the Baseline Version
in the postgraduate group should have comprehended the Baseline Version as well as those of
the Macro Version at the expense of spending more time decoding discourse structures. In other
words, if they had concentrated on the thread of ideas of the text along the linear structure of
the paper, they should have read efficiently and smoothly even though no markers were used
43

Vol.2, No.2, pp.27-49, June 2014
to signal the transition of the ideas. On the contrary, it turned out that although the subjects of
the Baseline Version in the postgraduate group had spent more time reading their paper, they
failed to understand it better than those who read the Macro Version. The only explanation is
that the subjects who read the Macro Version in this group had benefited much more from the
use of macro-markers. It is worth noting how the investigated readers of the Macro Version
could read better and faster than those of the Baseline Version. Since the subjects of the Macro
Version also read their paper word by word, then how could we explain the role played by the
macro-markers in helping the readers read the Macro Version better and faster?
The first explanation is that macro-markers are usually made up of common1y-used words,
with which the readers were quite familiar and were therefore easy to decode. In this sense,
these macro-markers played their role in signaling the major transitions and conveying the
emphases expressed in the text. Hence, the investigated students were able to employ these
signals to successfully carry out a triple reading process (Dr. Yorio, 197l ), i.e. the storage of
the past cues, prediction of the future cues, and associations between the two. This triple
reading process strengthened the readers ability to recall the facts they had just learned from
the paper. As they efficiently received many productive language cues to determine the
information of the text, they could move faster from one point to another as well.
Miller explains this phenomenon in this way: "It has been established that there is a severe limit
to the amount of information that we are able to receive, process and remember. The reader,
therefore, does not use all the information on the page, but rather, must select the most
productive language cues in determining the message of the writer "(Dr.Miller,
1975).Goodman also points out: "Reading is a selective process. It involves partial use of
available minimal language cues selected from perceptual input on the basis of the readers
expectation. As this partial information is processed, tentative decisions are made to be
confirmed, rejected or refined as reading progresses. "(Dr. Goodman, 1972)Therefore, there is
no wonder that, with the help of macro-markers, the investigated students could read the Macro
Version better and faster.
The third point for discussion is concerned with why the Micro-Macro Version did not show
any superiority over the Baseline Version in terms of helping the readers increase the reading
comprehension and speed. In the first place, the further addition of micro-markers to the macromarkers achieved no better results. Rather, it cost the readers more time in reading more
markers, and yet getting no further information valuable in any sense. This only distracted the
readers from observing and decoding the macro-markers. In the second place, owing to the
possibility of not knowing the meaning of some micro-markers and too much relying on the
word-by-word reading habit, the readers had often to stop now and then to analyze the
relationships between the unfamiliar transitional units introduced by those unfamiliar micromarkers and macro- markers. Consequently, they could not go on reading smoothly. Moreover,
their frequent hesitations before these troubles inevitably made them lose the thread of thinking
and memorization. That is, the past cues were forgotten, the present cues were lost, and there
44

Vol.2, No.2, pp.27-49, June 2014
was no way to predict what was coming next and to make associations between the storage of
past cues and the prediction of future cues. Dr. Omaggios investigations support this
explanation. As she describes, "Rather than recalling cues with which they are familiar, ESL
[English as a second language] readers are forced to recall cues that they either do not know at
all or know imperfectly. Because of this, the readers will forget those cues much faster than
they would in their native language. They must simultaneously predict future cues and make
associations with past cues, a slow and painful process in the second language for many
inexperienced learners. "(Dr. Omaggio, 1987) This study proves that it is also a slow and
painful process in the foreign language for many inexperienced learners.
It is evident, therefore, that the addition of micro-markers tends to decrease the role and
functions of macro-markers when these two kinds of markers go hand in hand. Hence, in this
survey the Micro-Macro Version showed no superiority over the Baseline, the Micro, and the
Macro Versions in terms of helping accelerate the reading comprehension and speed.
The fourth point to be discussed is why macro-markers could not lead the subjects of the
postgraduate group to their better recall of the ideas and facts of the material and to their faster
reading than micro-markers. If we consider the role of discourse markers, macro-markers, as a
rule, should be regarded as a better role to play in promoting efficient reading than micromarkers. But, in this study, there were no significant differences between the Macro Version
and the Micro Version.
One explanation is that some micro-markers used in the versions of the test paper were of
relatively more semantic value in conveying the information and could often reveal some
signals about the organization of the text. Since the subjects in the postgraduate group were at
the advanced level of English proficiency, those who read the Micro version must have
benefited from micro markers by efficiently decoding the information implied by these
markers.
Perhaps Dr. Clarke's following remarks might give another satisfactory answer: "ESL students
should be made aware of the strong tendency in English for linear argumentation. Unlike many
other languages, contiguous English sentences often imply causation or chronological
sequence of events. Therefore, it is necessary to emphasize the arrangement of ideas as an
important clue in deciphering the overall meaning of a text. "(Dr. Clarke, 1986)Therefore, even
if micro-markers signal less discourse information than macro-markers, the linear
argumentation itself in the paper might have better helped the advanced readers in reading the
Micro Version.
Dr. Halliday and Dr. Hasan also support this explanation as they point out: "Although the
cohesion is achieved through the conjunctive expression, it is the underlying semantic relation
that actually has the cohesive power. This explains how it is that we are often prepared to
recognize the presence of a relation of this kind even when it is not expressed overtly at all.
45

Vol.2, No.2, pp.27-49, June 2014
We are prepared to supply it for ourselves, and thus to assume that there is cohesion even
though it has not been explicitly demonstrated. "(Dr. Halliday and Dr. Hasan, 1976. p. 226)
Clearly, in this study both the underlying semantic relation of the Micro Version and the
subjects command of medical knowledge helped the readers of the Micro Version a lot in
terms of decoding the information contained in the text and therefore they were able to get the
same recall score at the same speed as those readers of the Macro version. By comparison, the
readers of the Micro Version in the undergraduate group were not familiar with the semantic
value of micro markers and hence failed to understand the underlying semantic transition of
the Micro Version. Consequently, they could not comprehend the Micro Version as well as
those who read the Macro Version. Nor could they read as fast as those who read the Macro
Version.
CONCLUSION
Attempts have been made in this survey to study the effects of discourse markers on the reading
comprehension and the reading speed. On the basis of this survey, the following conclusions
can be drawn.
1. Macro-markers help the readers who are either at the intermediate level or advanced level
better comprehend the paper because they add enough content or signals to make the
subsequent information more salient or meaningful.
2. Also macro-markers help them increase their reading speed because they are usually made
by commonly used words and can be easily decoded.
3. The apt use of micro-markers help readers read the paper efficiently. But this does not mean
that the more micro-markers are used, the better comprehension and the faster reading readers
can achieve. Instead, many micro-markers added to the paper may hinder readers in
comprehending it and even prevent them from reading more smoothly.
4. The combined use of both micro and macro markers in the version appears not to help the
readers comprehend best and read fastest as this study hypothesizes. The reason might be that
the combined use of micro-markers and macro-markers make the entire paper merely appear
less well organized.
The implications of this study are supposed to be useful for both teachers and learners of
English.
To teachers of English
In order to develop their students skills in gaining better reading comprehension and faster
reading speed, teachers should make their students be aware that both the formal schema and
content schema can enhance their reading comprehension process. While teaching the formal
schema, they should stress the functions of macro-markers because the knowledge of these
markers can enhance students reading skills so as to comprehend reading materials better and
read them faster. Furthermore, in the process of teaching those students whose English
proficiency is at the intermediate level, emphasis should also be placed on the teaching of
micro-markers, which commonly used in academic papers that abound in analyses and complex
theories. As far as English writing course is concerned, students should be encouraged to use
46

Vol.2, No.2, pp.27-49, June 2014
discourse markers more effectively because the apt use of them can help students achieve
cohesion and coherence in their writing. However, English teachers must warn their students
against any lavish use of micro-markers because it may destroy the overall coherence of their
writing.
To English learners
It is important that more attention should be paid to the ongoing "thread" of the information
conveyed in the paper by carefully decoding the discourse structures signaled by discourse
markers, particularly micro markers. And the knowledge of discourse markers is especially
beneficial when students read academic papers, which are characteristic of the use of a number
of discourse markers.
To conclude, discourse analysis should be an important teaching content in Chinas English
reading class of today. And students command of the knowledge of discourse markers can
assist them to comprehend English papers better and read them faster.
NOTES
1. "Discourse" refers to formal, extended, spoken or written treatment of a given subject. It has
several properties: it forms structures, conveys meanings and accomplishes actions. For more
information, see "Discourse Markers" by Schiifrin (1996).
2. Yuan (1982) once made a study in the U.S. on Chinese scientists difficulties in
comprehending English science lecture; and he found: "In general, the subjects were rather
weak at paying attention to the sequence of the lecture because of their neglect of the logical
connectors of sequence and their lack of recognition of transition from one main idea to
another. Besides, they paid more attention to decoding the speech sentence by sentence than to
extracting the science information from the lecture through understanding the rhetorical nature
and functions of both textual and lecture discourse."
3 . SPSS/PC (Statistical Program for Social Science -- Personal Computer) is the most powerful
software package for microcomputer data management and analysis in the world, which is
produced and distributed by SPSS Inc.
4." N " refers to number of the subjects.
5. "s.d." refers to "standard deviation ". The standard deviation is the most widely used
statistical measure of the spread or dispersion of a set of data. It is the positive square root of
the variance. In this survey the standard deviation is smaller, which means the differences of
the subjects in their reading comprehension and speed are not significant. And the differences
in the means must have resulted from the paper versions.
6. "SS" refers to "sum of squares".
7. "d f refers to "degree of freedom".
8. "MS" refers to " mean square ".
9. "F" refers to calculated statistics.
10. "P" refers to probability value.
11. Q value refers to calculated statistics;
12. a refers to the number of groups;
13. v refers to degree of freedom within groups;
47

Vol.2, No.2, pp.27-49, June 2014
14. q refers to significant level;

REFERENCES
1. Bails. D.C. et a1., "Urinary Sodium in the Evaluation of Hyperchloremic Metabolic
Acidosis. (J) THE NEW ENGLAND JOURNAL OF MEDICINE. Jan. 15. 1987.
2. Brown. G. and George Yule, Discourse Analysis (M) Cambridge University Press,
Cambridge. 1983, 2003
3. Carrel, P.L. Cohesion is not coherence. (J) TESOL Quarterly 16: pp. 479-88, 1982
4. Chaudron, Craig, and Jack Richards. "The effect of Discourse Markers on the
Comprehension of Lectures" (J) Applied Linguistics ; Vol. 7 (2). 1986.
5. Clarke. M.A. (M) Methodology in TESOL 1986
6. Cook. J. R. S. A Communicative Approach to the Analysis of Extended Monologue
Discourse and its Relevance to the Development of Teaching Materials for ESP. Thesis
University of Edinburgh, Scotland, 1975,
7. Goodin, G. "Discourse Analysis and the Art of Coherence." (J) College English Vol. 44 (1)
1982.
8. Goodman, K.S. "Reading: A Psych-linguistic Guessing Game." In Larry Harris and Carl
Smith, eds. Individualizing Reading Instruction: A Reader. N.Y. 1972
9. Guth, Hans P Words and Ideas (M) Wadsworth 5th ed. 1980
10. Dr. Halliday, M., and Dr. Hasan Cohesion In English (M) Longman May3, 1976
11. Henning. G. et al. Listening Reca1lA Listening Comprehension Test for Low
Proficiency Learners. A paper presented at AILA Lund. 9 August. 1981
12. Hiller, J.H. et al, A computer investigation of verbal characteristics of effective classroom
lecturing." 66 pp. 1968. Paper read at the American Educational Research Association.
13. Jean Davies, Raping and making love are different concepts: so are killing and voluntary
euthanasia (M) Journal of Medical Ethics 1988, 14, 148-149
14. Kintsch, W. ,and van Dijk. The Representation of Meaning in Memory (M) John Wiley
& Sons Inc. 1975
15. Miller, G. A. The Psychology of Communication, (M), HARPER ANDROW, 1975
16. Omaggio. A.C. Teaching Language In Context (M)
Heinle & Heinle; 1987, 1993, 2000
17. Quirk, R. et al. Grammar of Contemporary English (M) Longman, 1972
18. Rumelhart, D.D. "Schemata: the building blocks of Cognition." Comprehension and
Teaching: Research Reviews, ed. by J. T. Guthrie, Newark Delaware, International
Reading Association 1981
19. Schiffrin, Deborah (M) Discourse Markers Cambridge University Press 1988.
20. Sloan, Gary. "The Frequency of Transitional Markers in Discourse Prose." (J) Applied
Linguistics Vol. 7 (1) 1986
21. Yorio, Garlos A. "Some Sources of Reading Problems for Foreign Language Learners."
(J)Language Learning, Vol. 21, Issue 1, Pages 107115, 1971
48

Vol.2, No.2, pp.27-49, June 2014
APPENDIX A Micro-markers and Macro-markers

Micro-markers
Macro-markers
Furthermore
What is more,
In addition
What is most important
Then
It is evident that
Moreover
The truth is that..
Also
One important point is that
And
It must be emphasized that
It is of note that
It means that
But
Nevertheless;
However;
Nonetheless
In spite of this,
So
Thus;
Hence;
Therefore;
Consequently
This is because
On the basis of this situation..
It is conclude that..
For example
Similarly
is an example;
Another example is that
49

The Effects of Discourse Markers On The Reading Comprehension and Speed of Chinese Learners of English

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

The Effects of Discourse Markers On The Reading Comprehension and Speed of Chinese Learners of English

Caricato da

Copyright:

Formati disponibili

International Journal of English Language and Linguistics Studies

Vol.2, No.2, pp.27-49, June 2014